Improve usability of the Regexp format

This commit is contained in:
Alexey Milovidov 2024-05-15 04:49:11 +02:00
parent 96e858480e
commit 89ef4e9b53
4 changed files with 19 additions and 3 deletions

View File

@ -13,10 +13,14 @@ namespace ErrorCodes
{
extern const int INCORRECT_DATA;
extern const int LOGICAL_ERROR;
extern const int BAD_ARGUMENTS;
}
RegexpFieldExtractor::RegexpFieldExtractor(const FormatSettings & format_settings) : regexp(format_settings.regexp.regexp), skip_unmatched(format_settings.regexp.skip_unmatched)
RegexpFieldExtractor::RegexpFieldExtractor(const FormatSettings & format_settings) : regexp_str(format_settings.regexp.regexp), regexp(regexp_str), skip_unmatched(format_settings.regexp.skip_unmatched)
{
if (regexp_str.empty())
throw Exception(ErrorCodes::BAD_ARGUMENTS, "The regular expression is not set for the `Regexp` format. It requires setting the value of the `format_regexp` setting.");
size_t fields_count = regexp.NumberOfCapturingGroups();
matched_fields.resize(fields_count);
re2_arguments.resize(fields_count);
@ -58,8 +62,8 @@ bool RegexpFieldExtractor::parseRow(PeekableReadBuffer & buf)
static_cast<int>(re2_arguments_ptrs.size()));
if (!match && !skip_unmatched)
throw Exception(ErrorCodes::INCORRECT_DATA, "Line \"{}\" doesn't match the regexp.",
std::string(buf.position(), line_to_match));
throw Exception(ErrorCodes::INCORRECT_DATA, "Line \"{}\" doesn't match the regexp: `{}`",
std::string(buf.position(), line_to_match), regexp_str);
buf.position() += line_size;
if (!buf.eof() && !checkChar('\n', buf))

View File

@ -31,6 +31,7 @@ public:
size_t getNumberOfGroups() const { return regexp.NumberOfCapturingGroups(); }
private:
String regexp_str;
const re2::RE2 regexp;
// The vector of fields extracted from line using regexp.
std::vector<std::string_view> matched_fields;

View File

@ -0,0 +1,2 @@
regular expression is not set
`Upyachka`

View File

@ -0,0 +1,9 @@
#!/usr/bin/env bash
# Tags: no-fasttest, no-parallel, no-ordinary-database, long
CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
# shellcheck source=../shell_config.sh
. "$CUR_DIR"/../shell_config.sh
${CLICKHOUSE_LOCAL} --query "SELECT * FROM format(Regexp, 's String', 'Hello')" 2>&1 | grep -o -F 'regular expression is not set'
${CLICKHOUSE_LOCAL} --query "SELECT * FROM format(Regexp, 's String', 'Hello') SETTINGS format_regexp = 'Upyachka'" 2>&1 | grep -o -F '`Upyachka`'